Determinants of Arab Public Opinion on Foreign Relations

Determinants of Arab Public Opinion onForeign Relations

PETER A. FURIA

Wake Forest University

RUSSELL E. LUCAS

Florida International University

Using Zogby International polling data from seven different Arab na-tions (Egypt, Jordan, Kuwait, Lebanon, Morocco, Saudi Arabia, and theUnited Arab Emirates) this paper offers a societal-level quantitativeanalysis (N ¼ 91 dyads) of the determinants of Arab public opinion to-ward 13 different non-Arab countries (Canada, China, France, Germa-ny, India, Iran, Israel, Japan, Pakistan, Russia, Turkey, the UnitedKingdom, and the United States). We first explore whether Arabpublic opinion toward these countries is predicted by general ‘‘realist,’’‘‘liberal,’’ ‘‘Marxist,’’ and/or ‘‘cultural’’ hypotheses suggested in the IR/foreign policy literature. After finding few statistically significant rela-tionships among these variables, we present evidence that Arab publicsevaluate non-Arab countries on the basis of those countries’ specificforeign policy behaviors throughout the wider Middle East (e.g., espe-cially those behaviors affecting Palestine and Iraq). Noting that theseevaluations occur in the context of competing identity frames, we pro-visionally link Arab publics’ concerns with ‘‘regional’’ matters to thehigh salience of ‘‘Arabist’’ identity among respondents to the Zogbysurvey.

Particularly since September 11, 2001, American political commentators havespeculated widely on the sources of apparent Arab resentment toward the UnitedStates (Kifner 2001; Friedman 2002a, 2002b, 2003; Khalidi 2002; Kristof 2002;Sardar & Davies 2002; Wright 2002; Telhami 2003a; Waterbury 2003). In so doing,these observers sometimes cite unique aspects of U.S. policy, including, first andforemost, the longstanding special relationship between the U.S. and Israel(Telhami 2001). At other times, however, the commentators echo a variety of moregeneralizable hypotheses about enmity and rivalry familiar to scholars of interna-tional relations, political psychology and foreign policy opinion (Mearsheimer1990; Russett 1993; Huntington 1993, 1996; Holsti 1996; Sulfaro and Crislip 1997;Russett and Oneal 2001; Krauthammer 2002; Roger 2002; Wallerstein 2002; Joffe2003; Kagan 2003). In this paper, we first offer a preliminary test of six of thesesorts of hypotheses, examining Arab attitudes toward the United States in the

Authors’ note: The authors are equally responsible for the contents of this paper. The authors would like to thank

Zackary Shipley for his assistance and the editors and anonymous reviewers of International Studies Quarterly fortheir helpful comments.

r 2006 International Studies Association.Published by Blackwell Publishing, 350 Main Street, Malden, MA 02148, USA, and 9600 Garsington Road, Oxford OX4 2DQ, UK.

International Studies Quarterly (2006) 50, 585–605

context of Arab attitudes toward foreign countries more generally. Some of thesehypotheses pertain to characteristics of the Arab countries themselves, many othersto characteristics of the countries that may be objects of Arab resentment, and therest to dyadic relationships between the two. We conclude that none of these generaltheories of enmity and rivalry finds support in our data.

We argue instead that ‘‘middle-range’’ explanations of Arab opinion toward theoutside world appear most persuasive and, specifically, that Arab publics evaluatenon-Arab countries based in large part on their relatively recent foreign policyactions throughout the Middle East. Although this finding, that specific foreignpolicy actions drive opinion, is broadly in keeping with Hans Morgenthau’s tra-ditional view that the major determinants of foreign policy opinion are ‘‘national,’’the situation in Arab countries is complicated by competing visions of what the mostrelevant national community is. We explore how multiple identity frames competewithin the Arab world, including Arab nationalism, more country-centered nation-alisms, and Islamist identifications. As the Zogby poll finds a relatively strong lean-ing toward the ‘‘Arabist’’ identity frame among respondents, we link the prevalenceof this type of identity to consensus among Arab publics regarding the behavior ofnon-Arab countries in the wider Middle East and, in particular, their policiesregarding Palestine and Iraq. We do, however, find variation among differentArab societies in their evaluations of certain non-Arab countries asked about in theZogby study, suggesting that Arab attitudes on foreign relations are by no meansmonolithic.

Method

We base our analysis on the Zogby International ‘‘Arab Values Survey’’ carried outin March and April of 2002 (Zogby 2002:61). The survey was conducted using face-to-face interviews in what we henceforth refer to as seven Arab ‘‘subject’’ countries.Specifically, 600 adults per country were surveyed in Egypt, Jordan and SaudiArabia, and 400 adults per country were surveyed in Morocco, Lebanon, Kuwaitand the United Arab Emirates. After dividing each county into geographic regions,participants within each region were selected at random (Zogby 2002:99). Zogbydoes, however, concede that the regional sampling frames themselves tended toover-represent urban populations and, in turn, ‘‘those [individuals] more likely tobe literate with more access to both satellite televisions and the Internet’’ (Zogby2002:99). Although we thus caution that members of the ‘‘attentive’’ public aresomewhat overrepresented in these data, this is of course often the case with surveydata from the Global South (i.e., if not due to coverage errors induced by samplingframes, then due to high levels of non-response among uneducated rural re-spondents). All in all then, we are about equally confident in the quality of the datacollected by Zogby as we are in that collected by other internationally respectedorganizations operating in the Global South (e.g., Gallup, Pew, and the WorldValues Study Group).1

The Zogby survey contains many questions pertaining to personal as well aspolitical matters, but the primary item that we analyze concerns Arab attitudestoward 13 non-Arab ‘‘object’’ countries. Specifically, respondents in each of the‘‘subject’’ countries were asked to answer the following question:

1 The margin of error for the country surveys ranges between � 4–6%. Although Zogby only surveys seven of18 Arab countries, the sample draws from over 50% of the population of the Arab world. Despite the authoritarian

nature of Arab governments, we find no empirical evidence that respondents failed to state their true preferences onforeign relations. Rather, the high degree of in-country variation in attitudes on foreign relations, as well as the clearwillingness of respondents to offer negative evaluations of countries with which their own governments possesssecurity ties (see below), lead us to believe that respondents felt little or no state pressure to answer the questionshere analyzed in a particular way.

Determinants of Arab Public Opinion on Foreign Relations586

Interviewer: ‘‘I will read you a list of countries. Please tell me if your overallimpression of each is either very favorable, somewhat favorable, somewhat un-favorable or very unfavorable, or if you are not familiar enough to form ajudgment.’’

The object countries in the Zogby study were chosen on the basis of their regionaland/or global political significance. They are Russia, China, the United States,France, India, Israel, Pakistan, Iran, Japan, Turkey, Canada, Germany, and theU.K. Although some might prefer a random selection of object countries, thecountries do vary in significant ways, both in terms of the predictors of Arab opin-ion that we examine and in terms of that opinion itself (i.e., in the extent to whichthey are positively regarded by Arab publics).

Specifically, the dependent variable in our analysis, AFFECT, represents thepercentage of respondents in a given subject country that express either ‘‘very’’ or‘‘somewhat’’ favorable feelings toward a given object country. We compute thisvariable in regard to 91 different subject country/object country ‘‘dyads’’ (e.g.,‘‘Lebanese attitudes toward France’’) based on the fact that Zogby conducted sur-veys in seven Arab ‘‘subject’’ countries and asked about attitudes toward 13 foreign‘‘object’’ countries in each of these surveys (i.e., 7 � 13 ¼ 91). Subject country affecttoward object countries varies from a low of 2 (representing chilly Kuwaiti andEmirati attitudes toward Israel) to a high of 79 (representing warm Kuwaiti atti-tudes toward Iran).2 The mean affect score for all 13 object countries across theseven subject countries surveyed is 43.2 with a standard deviation of 19.4. We nowevaluate six ‘‘general’’ theories of threat perception, enmity, and rivalry as predic-tors of Arab attitudes on foreign relations. Although these theories are often com-peting, all are grounded in either the ‘‘realist,’’ ‘‘liberal,’’ ‘‘Marxist,’’ and/or‘‘cultural’’ schools of IR scholarship.

General Hypotheses

Throughout the Cold War, most realist scholars of international politics assumedthat analysts could ignore the extent to which publics participate in the enmitiesand rivalries of their political leaders.3 Noting an apparent rise in ‘‘hyper-nation-alism’’ and ‘‘mass hate’’ after the Cold War, however, many realists have subse-quently changed their minds. Perhaps due to the relative infrequency of interstatewar in the early post-Cold War era, proponents of the ‘‘neorealist’’ view that bal-ances of power promote peace have seemed particularly eager to ‘‘bring publicopinion back in.’’ Beginning with John Mearsheimer, realists who see the currentinternational distribution of power as unstable have argued that even if powerpreponderance promotes peace in the short term, imbalances of power betweencountries breed long-lasting hatreds and resentments that may ultimately lead toviolent conflict (Mearsheimer 1990; Roger 2002; Joffe 2003; Kagan 2003). In otherwords, ‘‘weak countries will always resent and seek to topple the strong.’’ Incontrast, ‘‘power transition theorists,’’ who see power preponderance as stability-enhancing, suggest that only when weaker countries begin to perceive strongcountries as weakening is the door opened to military challenges (Blainey 1973;Organski and Kugler 1980; Gilpin 1981; Krauthammer 2002). Put another way,‘‘the weak will always respect the strong.’’ By examining how, if at all, a state’smilitary capabilities impact perceptions of that country abroad, we can simultane-

2 Notably, these outliers are only partly suggestive of patterns in the sample as a whole.3 As Kenneth Waltz argued in 1959, for example, the fact that popular ‘‘peace wishes’’ may have run strong

among both the Soviet and American peoples during the Cold War seemed to do little to reduce Cold War–erasecurity competition between the U.S. and USSR (Waltz 1959:3). Other Cold War–era realists such as HansMorgenthau and E.H. Carr were of course more interested in public opinion, yet their views were seemingly lessinfluential on scholarship of the period.

PETER A. FURIA AND RUSSELL E. LUCAS 587

ously and systematically test both of these claims. We formalize this test via Hy-pothesis 1.

Hypothesis 1: Positive attitudes expressed by citizens of State A toward State B aresystematically related to the distribution of military capabilities between the two states.

We operationalize Hypothesis 1 in multiple ways. Typically, IR scholars calculatea ‘‘capabilities ratio’’ for a dyad by taking the six-item ‘‘capabilities score’’ for thestronger of two countries in the Correlates of War (COW) data sets and dividing itby the capabilities score of the weaker country (Singer, Bremer, and Stuckey 1972).4

We likewise calculate this conventional indicator of the ‘‘capabilities ratio’’ for eachdyad in the traditional manner, using data on capabilities updated through 2001by the authors of the EUGene statistical software package (Bennett and Stam2000). Note, however, that Hypothesis 1 technically refers to a ‘‘directed-dyad’’relationship in regard to public attitudes on the part of the weaker member of adyad. Thus, although our Arab subject countries have lower capabilities scores thanthe object countries that their publics are asked about in 89 out of 91 cases (the onlytwo exceptions being the Egypt–Israel dyad and the Saudi Arabia–Israel dyad), wealso calculate a directed capabilities measure of ‘‘object-country preponderance,’’which consists of the ratio of the object-country’s capabilities score to that of thesubject country. Finally, we create a variable measuring ‘‘object-country capability’’in absolute terms, in order to test for the possibility that strong countries will berespected or resented by subject countries irrespective of the ratio of power be-tween them.

To note that realists are increasingly attentive to societal-level nationalism andresentment is not to deny that many realists see societal enmity as an ‘‘epiphe-nomenal’’ product of the strategic action of state elites. In this vein, a secondbroadly realist hypothesis posits that malleable publics can be made to expresssuitably favorable attitudes toward any and all foreign countries with which stateleaders determine that their security interests compel them to ally. To cite only oneparticularly famous example, Henry Kissinger and other realists have celebratedFDR’s adroitness in responding to shifting strategic needs by manipulating U.S.public opinion toward Germany and the U.S.S.R. in various directions over thecourse of his presidency (Kissinger 1994). On the other hand, the claim that the‘‘Arab Street’’ has proven highly resistant to similar attempts on the part of Arabstate elites is no less familiar. Indeed, some commentators go so far as to suggestthat insofar as Arab publics view their own governments as oppressive, they willsystematically resent precisely those countries with which their governments possesssecurity ties (Wright 2002). We formalize both of these competing claims via Hy-pothesis 2.

Hypothesis 2: Positive attitudes expressed by citizens of State A toward State B aresystematically related to security ties between the two states.

To operationalize Hypothesis 2, we again utilize the EUGene software package tocreate a dummy variable for the presence of any of the various types of alliancescoded in the most recent update of the Correlates of War alliance data (Bennett andStam 2000). Perhaps tellingly, however, none of the 91 dyads in our data set in-volves two states coded as having any sort of alliance by COW. (Without furtheranalysis, we can thus reject the hypothesis that variation in AFFECT is explained bysecurity ties so defined.) In order to test Hypothesis 2 a bit more generously,however, we also created a dummy variable ‘‘security ties’’ that we coded ‘‘1’’ for the

4 The six equally weighted components of a country’s capabilities or CINC ‘‘Composite Index of MilitaryCapabilities’’ score are total population, urban population, energy consumption, iron and steel production, militaryexpenditures, and military personnel (Singer et al. 1972).


20 dyads for which we found government and/or press reports of informal agree-ments and/or joint security exercises in the three years before the survey.

‘‘Liberal’’ scholars of international politics have proposed numerous hypothesesabout the relationship of public opinion to foreign policy. For the purposes of thisparticular study, though, only one such hypothesis seems directly relevant.5 Name-ly, many liberals suggest that through its tendency to promote amity or ‘‘fellow-feeling’’ among societies, economic interdependence between states promotesdyadic peace. Although some liberals view the profits from trade and investmentaccruing to capitalist elites as alone sufficient to pacify dyadic relations, the mostprominent among them argue that trade exercises much of its pacifying effectbecause it leads to ‘‘mutual respect’’ among trading societies (see Doyle 1983; Rus-sett and Oneal 2001). We formalize this liberal claim via Hypothesis 3.6

Hypothesis 3: Positive attitudes expressed by citizens of State A toward State B constitute amark-up on levels of economic interdependence between the two states.

The most commonly utilized COW data set on dyadic trade likewise presentsproblems for the operationalization of Hypothesis 3, in that, thus far, it has onlybeen updated through 1992.7 Luckily, our relatively small sample of 91 dyadsmakes it possible for us to compute 2001 trade statistics for each of our pairs ofcountries. Specifically, we operationalize Hypothesis 3 by calculating two variables.The first variable, ‘‘gross dyadic trade,’’ consists of the sum of exports and importsbetween the subject country and the object country as reported in the 2002 IMFDirection of Trade Statistics Yearbook (IMF 2002). The second variable, ‘‘Gross Do-mestic Product (GDP)-adjusted dyadic trade,’’ is calculated by dividing this figureby the relevant subject country’s GDP.8

‘‘Marxist’’ and ‘‘dependency school’’ scholars of international relations also sug-gest at least one hypothesis relevant to the present analysis.9 Contrary to liberals,Marxists argue that asymmetric trade between countries may lead to resentmentwithin a dyad insofar as one country within a dyad is disadvantaged by the terms ofthat trade. The claim that trade dependence fuels resentment of foreign exporterscan, of course, be applied to northern publics (e.g., it is often adduced in explainingthe American public’s resentment of Japan in the 1980s and 1990s), but Marxistsmost frequently articulate this claim in regard to ‘‘neo-colonial’’ dependencies suf-fered by countries in the Global South (Sardar and Davies 2002; Wallerstein 2002).With Hypothesis 4, we test the claim that trade dependence on the part of Arabcountries influences attitudes toward the foreign countries on which they depend.

Hypothesis 4: Negative attitudes expressed by citizens of state A toward state B constitute amark-up on A’s level of import dependence on B.

5 The relatively robust hypothesis that joint democracy promotes dyadic peace, for example, is not straight-forwardly related to our present study of attitudes in Arab subject countries (none of which was coded as democraticin either the ‘‘Polity’’ or ‘‘Freedom House’’ data sets as of 2002).

6 Foreign direct investment between societies (FDI) may also have pacifying effects, but data on FDI are muchless readily available than data on trade. See Russett and Oneal (2001) for an argument that due to the prominence

of intra-industry trade, trade data are nonetheless a fair proxy for overall economic links between societies. Asimultaneous relationship between economic links and positive affect is also possible (i.e., societies may trade morewith foreign societies because they regard them favorably), but as we find no link between trade and positive affectbelow, the direction of this (non)-relationship is of limited concern.

7 See Jones et al. (1996) as well as the ‘‘downloads’’ page of the Correlates of War project, http://cow2.la.psu.edu/.8 We also tested the effects of logged values of dyadic trade on AFFECT, but we found no statistically significant

results here either (analyses included in an earlier version of the manuscript available from the authors).9 Other Marxists hypothesize that ‘‘poor’’ countries of the Global South may be systematically hostile to ‘‘rich’’

countries of the Global North. This hypothesis can also be tested empirically, but, as it turns out, neither an object-country’s GDP nor its logged GDP nor its GNI per capita is significantly related to the attitudes toward that countryheld by Arab publics (analyses available from the authors).


Using the same IMF data as we do to operationalize Hypothesis 3, we measure asubject-country’s level of ‘‘import dependence’’ on an object country in two ways.First, we calculate a variable ‘‘GDP-adjusted imports’’ in a manner directly parallelto that which we use in calculating ‘‘GDP-adjusted dyadic trade’’ above. This var-iable allows us to distinguish economically significant export dependencies frominsignificant ones, but it fails to distinguish between dyads in which exports flowboth ways and dyads in which exports are one sided. Thus, we likewise calculate a‘‘dependency ratio’’ for each dyad, consisting of a subject country’s absolute level ofimports from a given object country divided by its absolute level of exports to thatsame country.

Although realism, liberalism, and Marxism together remain the dominant sourc-es of globally testable hypotheses on international relations, their inability to ex-plain Arab attitudes on foreign relations might not strike scholars such as SamuelHuntington as much of a surprise (Huntington 1993, 1996). By contrast, what wehere term the ‘‘cultural’’ school of IR scholarship has proposed numerous generalhypotheses that are often thought particularly pertinent to Arab and/or Muslimcountries. Last but not least, we examine two closely related hypotheses about Arabopinion on foreign relations suggested in the work of Huntington and his followers,both broadly associated with what he refers to as a ‘‘clash of civilizations.’’

In his initial 1993 article, Huntington was widely understood as arguing that‘‘civilizational’’ identity would expand worldwide, perhaps soon eclipsing allegianceto nation states (Huntington 1993). Although few regions of the world (including,emphatically, Europe) have exhibited this trend to date, Mark Tessler and JodiNachtwey (1998) have, intriguingly, found relatively high levels of allegiance to theIslamic umma in Arab countries. In subsequent writings, however, Huntington hasmade clear that he does not think that nation states and patriotic allegiances areweakening, but only that ‘‘kin-countries’’ from a given civilization will identify morestrongly with each other than with other sovereign countries. We formalize thisclaim via Hypothesis 5.

Hypothesis 5: Positive attitudes expressed by citizens of State A toward State B constitute amark-up on the existence of a shared ‘‘civilization’’ between the two states.

In addition, Huntington’s work suggests a hypothesis about Arab (or, moreproperly, Muslim) attitudes on foreign relations that is somewhat less ‘‘general’’than the other hypotheses that we have mentioned thus far. Namely, he argues thatbeyond the fact that Muslim and Western countries feel no special affinity for eachother, there is an especially pronounced tendency for countries from these twocivilizations to ‘‘clash.’’ Although this hypothesis has proved extremely controver-sial, it is worth pointing out that it does find some support in Western public opiniondata.10 At present, however, we are interested in determining whether a parallelpattern obtains among the Arab publics surveyed by Zogby, as suggested by Hy-pothesis 6.

Hypothesis 6: Citizens of ‘‘Muslim’’ countries will express systematically negative attitudestoward ‘‘Western’’ countries.

In order to operationalize these two ‘‘civilizational’’ hypotheses, we initially followHuntington’s own guidelines for assigning countries membership in a civilizationbased on cultural and, in particular, religious attributes of the majority of thecountry’s population (Huntington 1993, 1996). Specifically, in order to operation-

10 For example, U.S. public attitudes in the 1994–2002 Chicago Council on Foreign Relations ‘‘American PublicOpinion and U.S. Foreign Policy’’ surveys (CCFR 1995, 1999, 2003) suggest a highly statistically significant tendencyto give less warm ‘‘thermometer ratings’’ to foreign countries with a majority Muslim population. The strength ofthis bivariate tendency ranges from r ¼ � .579 ( po.002) in 1998, to r ¼ � .624 ( po.001) in the post-9/11 surveyof 2002.


alize Hypothesis 5, which holds that the publics of our majority-Muslim subjectcountries will feel a special affinity for majority-Muslim object countries, we create adichotomous variable ‘‘Muslim–Muslim dyad’’ and code it ‘‘1’’ for dyads involvingthe three out of 13 object countries (Iran, Pakistan, and Turkey) that Huntingtonhimself categorizes as ‘‘Muslim’’ on the basis of Islam’s being the most widelypracticed faith therein. Similarly, in order to operationalize Hypothesis 6, whichholds that the publics of our ‘‘Muslim’’ subject countries will feel a special hostilitytoward ‘‘Western’’ object countries, we create a dichotomous variable ‘‘Muslim–Western dyad,’’ which we code ‘‘1’’ for dyads involving the five out of thirteenobject countries (Canada, France, Germany, the U.K., and the U.S.) that Hunting-ton himself categorizes as Western (i.e., majority ‘‘Western-Christian’’). Notably,although Huntington himself categorizes Israel as a ‘‘non-Western’’ country (Hunt-ington 1996:91), Israel’s unpopularity in Arab countries makes our coding decisionin its regard far from trivial. Thus, although we adopt Huntington’s coding for ourbivariate analysis, our multivariate models examine the effects of operationalizingdyads involving Israel in multiple ways.

Results

Table 1 summarizes bivariate relationships between AFFECT and the multiple in-dependent variables discussed in accordance with each of our six hypotheses. Thesebivariate relationships suggest that, no matter how the relationships predicted byrealism, liberalism, Marxism, and ‘‘cultural IR’’ are operationalized, statisticallysignificant predictors of AFFECT are nowhere apparent. In turning to a morerigorous multivariate analysis, therefore, we choose to model those six indicatorsthat nonetheless cast their respective hypotheses in the most favorable possible light(with the exception of one variable whose introduction into our multivariate modelswould also introduce a problem with missing data).11

Specifically, as shown in Table 2 below, Models I–IV operationalize Hypothesis 1in terms of the absolute military capability of an object country; Hypothesis 2 interms of the informal security ties between countries; Hypothesis 3 in terms of theGDP-adjusted measure of trade within a dyad; Hypothesis 4 in terms of the GDP-adjusted level of object-country imports into a given subject country; Hypothesis 5 interms of a dummy variable coded ‘‘1’’ if the majority religious persuasion in bothsubject country and object country is Islam; and Hypothesis 6 in terms of a dummyvariable coded ‘‘1’’ if it involves a ‘‘Western’’ object country. (Models I–IV vary onlyin terms of dyads that are coded as ‘‘Muslim–Western’’ and/or are excluded fromthe analysis.)

Model I shows the results of a multivariate ordinary least squares (OLS) analysisof our six predictors of AFFECT across all dyads when, following Huntington,dyads involving Israel are not coded as ‘‘Muslim–Western’’ dyads. As we mightexpect given the negative bivariate findings in Table 1, none of the six predictors ofAFFECT shows up in this model as even borderline statistically significant. Joint R2

for the model is a low .041, so we now have somewhat more robust evidence thatrealism, liberalism, Marxism, and ‘‘cultural IR’’ rank as roughly equal failures whenit comes to predicting Arab attitudes on foreign relations. Model II diverges fromModel I only in that we recode the seven subject–object dyads involving Israel topositive values on the ‘‘Muslim–Western’’ variable. Ironically, it is only by violating

11 In regard to Hypothesis 4, we include the ‘‘GDP-adjusted imports from object country’’ instead of the onlyslightly more robust ‘‘imports/exports from object country’’ variable because including the latter would reduce our

sample size from 91 to 81. Although our bivariate results find no systematic support for Marxist suspicions about theresentment-inducing properties of asymmetric trade, therefore, it is worth noting that our multivariate models are abit less generous in their operationalization of Marxist hypotheses than of hypotheses drawn from other IR par-adigms. Note, moreover, that the negative relationship between ‘‘gross dyadic trade’’ and AFFECT is actually moreconsistent with Marxism than it is with liberalism.


Huntington’s own coding rule in regard to Israel that we uncover any evidence of aMuslim–Western ‘‘clash.’’ Whereas, in Model I, Arab publics actually appearedslightly positively disposed toward the West (albeit not significantly so), Model II’srecoding of a mere seven out of 91 dyads now produces a highly statistically sig-nificant tendency for Arabs to be negatively disposed toward Western countries. Asin Model I, the coefficients on the five predictors corresponding to Hypotheses I–V

TABLE 1. Bivariate Associations of AFFECT and Multiple Independent Variables Corresponding toHypotheses 1–6

Independent Variables N Pearson r p-ValueIncluded in

OLS Models?

Capabilities variables (H1)Capabilities ratio (non-directed) 91 .002 .998 NoCapabilities ratio (directed) 91 .005 .965 NoAbsolute capability of object 91 � .047 .655 YesAlliance variables (H2)COW alliance (all values ¼ 0) 91 NA NA NoSecurity tie 91 � .053 .618 YesLiberal trade variable (H3)Absolute dyadic trade 91 � .106 .315 NoGDP-adjusted dyadic trade 91 .012 .909 YesImport dependence variables (H4)Imports/exports from object country 81 .095 .398 NoGDP-adjusted imports from object country 91 .010 .923 Yes‘‘Kin countries’’ variable (H5)Muslim–Muslim dyad 91 .122 .247 YesCivilization ‘‘clash’’ variable (H6)Muslim–Western dyad 91 .047 .660 Yes

Note: p-values calculated using two-tailed tests of significance.

TABLE 2. Multivariate OLS Models Testing Realist, Liberal, Marxist, and Cultural Variables as Pre-dictors of AFFECT

PredictorModel I

Israel 6¼ WesternModel II

Israel ¼ Western

Model IIIIsrael Dyads

Deleted

Model IVUS & Israel

Dyads Deleted

(Intercept) 37.378nnn 51.493nnn 57.130nnn 45.993nnn

(4.652) (5.561) (4.853) (8.414)Object’s capability score 13.234 � 80.524 � 119.392nn 20.717

(53.470) (56.184) (48.556) (99.814)Informal security tie � 5.841 1.999 � 3.272 � 1.675

(5.801) (5.543) (4.826) (5.189)GDP-adjusted dyadic trade 36.247 � 66.080 � 26.558 13.767

(95.597) (92.871) (79.834) (89.802)GDP-adjusted imports(from object country to subject)

� 37.089 335.597 � 2.938 � 78.855(265.160) (248.467) (220.195) (240.219)

Muslim–Muslim dyad 9.723 � 4.905 � 7.751 .790(6.011) (6.543) (5.624) (7.795)

Muslim–Western dyad 7.777 � 16.110nnn � 5.725 3.337(5.951) (5.986) (5.330) (8.061)

N 91 91 84 77R2 .041 .101 .111 .005

Main entries in table are OLS regression coefficients; standard errors in parentheses below.p-values calculated using two-tailed tests of significance; n ¼ po.10; nn ¼ po.05; nnn ¼ po.01.GDP, Gross National Product; OLS, ordinary least squares..


remain non-significant, and, indeed, the sign on all five of these insignificant pre-dictors has changed.

The considerable differences between Models I and II suggest that a third ap-proach to the dyads involving Israel is in order. Although we have thus far beenhesitant to delete any dyads from our analysis, removing the Israel dyads seems alogical compromise between coding them as either a ‘‘0’’ (Model I) or a ‘‘1’’ (ModelII) on our variable testing for ‘‘Muslim–Western’’ enmity. The results of removingthese seven dyads are presented in Model III, which shows a negative but notstatistically significant coefficient for the variable corresponding to Hypothesis 6. Inother words, Hypothesis 6 again fails to be confirmed even given a more generousoperationalization of Huntington’s ‘‘clash’’ thesis than seems strictly warranted.

Intriguingly, however, removing the Israel dyads from the data set initially seemsto paint the ‘‘neorealist’’ variant of Hypothesis 1 in a much more favorable light.Setting aside Arab publics’ deep resentment of Israel (a country not coded as par-ticularly powerful by COW) Model III suggests that Arab publics are otherwisequite systematically resentful of ‘‘strong’’ countries, with object-country capabilityshowing up as a statistically significant predictor that alone explains over 6% of thevariation in AFFECT. Upon further analysis, however, even this apparently ‘‘gen-eral’’ finding seems entirely attributable to negative attitudes about a single foreigncountry: in this case, the United States. For, as is shown in Model IV, if we removenot only the seven dyads involving Israel but also the seven dyads involving theU.S. from the analysis, the relationship between AFFECT and object-country ca-pability turns insignificant and changes its sign.

What is perhaps most remarkable about Model IV is that once the U.S. and Israelare both removed from the analysis, our six independent predictors of Arab opin-ion jointly explain only one-half-of-one percent of the variance in AFFECT(R2 ¼ .005)! To be sure, Arab attitudes toward countries other than the U.S. andIsrael vary enormously, and we should not thus conclude that they are in any sense‘‘random.’’12 Rather, the findings of Model IV tend to bolster the conclusion al-ready suggested by the evidence of Models I–III; namely, Arab attitudes toward allcountriesFthe U.S. and Israel includedFare determined by factors other thanthose thus far specified by ‘‘general’’ IR theories.

In the next section of the paper, therefore, we take a closer look at the Zogby datain order to develop new hypotheses about what these undertheorized determinantsof Arab opinion might be. We find substantial empirical support for the familiar ifhardly scientific assertion that Arab attitudes toward foreign countries are drivenless by what those countries ‘‘are’’ than by what those countries ‘‘do.’’ We examinethe theoretical status of this popular claim, considering how and whether it can bequantitatively operationalized in the Middle East and beyond. In particular, wehighlight two major difficulties that any quantitative analysis of ‘‘behavioral’’ de-terminants of Arab opinion must confront. A first is that the myriad foreign policy‘‘deeds’’ that appear to be linked to Arab opinion are rarely accounted for inexisting IR data sets. A second, and more theoretically significant, difficulty is thatwe often find that Arab citizens of a given subject country evaluate a given objectcountry less on the basis of its behavior toward that subject country than on thebasis of its treatment of Arabs throughout the Middle East. Although remaining opento the possibility that this tendency will prove generalizable to other regions, weultimately suggest that Arab opinion is influenced by numerous contextual factorsthat would help to explain a possibility we think more likely: namely, that otherpublics fail to share Arabs’ high levels of concern with ‘‘regional’’ matters.

12 AFFECTscores for the 77 dyads not removed from Model IV via listwise deletion of missing data vary almost asmuch as those analyzed in the models including the U.S. and Israel (ranging from a low score of 11 for Emiratiattitudes toward Turkey to the aforementioned high score of 79 for Kuwaiti attitudes toward Iran).


Foreign Policy Behavior as a Determinant of Arab Opinion

The analysis above suggests the failure of ‘‘realist,’’ ‘‘liberal,’’ ‘‘Marxist,’’ and ‘‘cul-tural’’ IR theories to explain why Arab publics possess extremely positive feelingstoward certain foreign countries and extremely negative feelings toward others. Wewould by no means rule out the possibility that these IR theories may performbetter in other contexts, but in this paper, a comprehensive global test of suchtheories is not our aim.13 Rather, our present purpose is to explore the determi-nants of foreign policy opinion in Arab countries (countries in which such opinion iswidely regarded as being of particularly high substantive interest). Having system-atically demonstrated that existing theory offers no compelling explanation for thepatterns observed in our data, we now examine these data more closely.

Table 3 presents both the average AFFECT score received by each object country(i.e., across all subject countries) and the standard deviation from this average(reflecting the extent to which the seven Arab publics surveyed diverge in theiropinions of the object country in question). It is not surprising that Israel and theU.S.Fthe two countries whose inclusion or exclusion from the OLS analysis soaffected its resultsFhave the two lowest mean affect scores among the 13 objectcountries. Perhaps more surprising is that attitudes toward France not only aremore favorable than attitudes toward any other object country, but also that theyshow less variation across subject countries than do attitudes toward any othercountry except Israel. We now consider the possibility that it is particular foreignpolicy behaviors of individual object countries that are in fact driving perceptions ofthese countries throughout the Arab world.

Starting at the bottom of Table 3, it is clear that the seven Arab publics in ourstudy hardly diverge at all in their attitudes toward Israel. Indeed, there is no Arabsubject country in which more than 8% of the public views Israel favorably. To most

TABLE 3. Mean Arab Country AFFECT in Regard to Particular Object Countries

Object Country Mean AFFECT Score Across Subject Countries Standard Deviation

France 60.57 9.64Iran 59.00 12.42Japan 57.29 14.43China 51.57 16.68Germany 49.43 16.25Canada 49.14 16.44Pakistan 46.43 10.36India 41.14 10.40Russia 38.29 15.98Turkey 34.43 13.14UK 32.86 11.65USA 25.29 12.72Israel 4.71 2.56Overall 42.32 19.34

Computed from Zogby, 2002a: 61 (Table XXXVI ‘‘How Arabs View Other Countries’’).

13 There is as yet no trans-regional survey that asks respondents in multiple subject countries their opinionsabout multiple object countries. One can create a pooled data set based on single-country and single-region surveys(e.g., the Chicago Council on Foreign Relations studies and certain Eurobarometer and Latinobarometer studies),

but the significant differences between these surveys create numerous analytical problems. Having made a pre-liminary attempt to do so, Furia (1999) comes to a similar conclusion about the predictive power of most theories ofenmity, as do Bennett and Stam (2003) about most theories of interstate conflict. Namely, of all of the determinantsof dyadic amity and enmity proposed by IR theory, only joint democracy seems to have a statistically significant effecton the attitudes expressed by members of a given ‘‘subject’’ country toward members of a given ‘‘object’’ country.


experts on the region, the fact that these exceptionally low ratings of Israel stemdirectly from the ongoing Israeli–Palestinian conflict goes without saying. Israel’scontinued military occupation of the West Bank, and, at the time of the survey,Gaza, as well as the perception that Israel refuses to recognize its role in creatingthe Palestinian refugee problem both help to create deeply unfavorable attitudestoward Israel among Arabs. As one Saudi columnist sums it up, ‘‘It’s the occupation,stupid’’ (Khashoggi 2002). Yet if it is unsurprising that Israel’s occupation of Pal-estine would predict hostility to Israel in Palestine, the fact that Israel is resentedthroughout the Arab world presents IR scholars accustomed to analyzing interstatedyads with more of a puzzle. Were we, for example, to create a dummy variable foran object country’s recent military occupation of a subject country, our entire dataset would include only one positive value for the Israel–Lebanon dyad.

In point of fact, however, Israel’s foreign policy actions have rarely been per-ceived as directed against Palestinians alone. Rather, the 1948, 1956, 1967, and1973 Arab-Israeli wars and the two Palestinian uprisings are all events that havebeen felt throughout the Arab world. Palestine’s central place as a political issue forArabs has been long established, and many suggest that as an issue, it has helpedcreate the Arab ‘‘public sphere’’ (Kerr 1971; Korany 1993). As early as the 1930s,‘‘residents of the region were now defining themselves as Arabs, supporting Arabsin Palestine and expecting their governments to do the same’’ (Barnett 1998:72).Today as well, according to Telhami, ‘‘No other issue resonates with the public inthe Arab World, and many other parts of the Muslim world, more deeply thanPalestine’’ (Telhami 2002:96).

Returning to contemporary empirics, Zogby’s ‘‘Arab Values Survey’’ itselfpresents compelling evidence of Palestine’s place as a central issue for Arabs inall countries surveyed. Table 4 shows that, in absolute terms, the vast majority ofArabs in all seven subject countries rank ‘‘Palestine’’ and the ‘‘Rights of Palestin-ians’’ as important or very important political issues. In relative terms, the (mostlynon-Palestinian) Arabs surveyed ranked Palestine as the third most important in-ternational or domestic political issue, surpassed only by civil/political rights andhealth care in the respondents’ own states (Zogby 2002:33–34).

To be sure, events which occurred as the Zogby survey was being conducted inlate March and April 2002 may have increased the salience of the Palestinian issuein Arab opinion.14 At the time of the survey, Saudi Crown Prince Abdullah hadrecently presented a plan to the Arab Summit to restart Palestinian–Israeli peace

TABLE 4. The Importance of Palestine as a Political Issue

Subject Country

% Ranking Important or Very Important

Palestine Rights of Palestinians

Egypt 89 86Jordan 83 86Kuwait 90 85Lebanon 69 67Morocco 92 90Saudi Arabia 97 96UAE 90 92Mean 87 86

Zogby (2002a:34); Table XIII ‘‘Importance of Political Issues.’’

14 In order to supplement the Zogby survey, we also surveyed the Arab press for the first 5 months of 2002 fornews and opinion stories regarding the 13 object countries. We cite from these media sources to further support ourargument about Arab perceptions of particular object countries.


negotiations, which had stalled in October 2000 with the beginning of the secondPalestinian intifada. The plan called for the recognition of Israel by the Arab statesin return for a full Israeli withdrawal to its 1967 borders. However, after a wave ofPalestinian suicide bombings in Israeli cities, Israel reoccupied a number of Pal-estinian cities in the West Bank in ‘‘Operation Defensive Shield.’’ In the ensuingbattles, Yasir Arafat was almost killed, and large portions of the Jenin refugee campwere destroyed. Yet while this series of events certainly thrust the Palestinian–Israeli conflict to the forefront of Arab and international attention, other surveysconfirm that the high levels of sympathy for the Palestinians recorded by Zogby arenot atypical (see, e.g., Center for Strategic Studies, University of Jordan 2005:79–80).

Beyond its importance in shaping Arab public attitudes toward Israel, the find-ings of Table 4 suggest that the issue of Palestine is so important to Arabs that it mayserve as a ‘‘litmus test’’ for their evaluations of other countries as well. If this isindeed the case, then it would seem reasonable to expect that, besides Israel itself,the object country most widely perceived as failing this litmus test would be theUnited States (i.e., due to its exceptional levels of military, economic and diplomaticsupport for Israel). As Table 3 suggests, only a quarter of Arabs surveyed by Zogbyview the United States favorably, easily rendering the U.S. the second-least popularof the thirteen object countries asked about in the study. Yet while it is tempting tosuggest that negative affect for the U.S. stems solely from its support for Israel,evidence from other Zogby surveys suggests a more complicated picture.

In particular, evidence from a separate April 2002 Zogby International ‘‘Im-pressions of America’’ poll suggests that even before the 2003 U.S. invasion of Iraq,‘‘extra-military’’ U.S. actions in regard to that countryFabove all the U.S.-dom-inated sanctions regimeFwere exercising a large negative effect on America’s im-age in the Arab world.15 Specifically, as indicated in Table 5, favorable views of theU.S. increased by nearly 50% in all countries except Kuwait when respondents wereasked about their potential attitudes toward the U.S. if it were to end sanctions onIraq. This is especially direct evidence that Arab publics in a given subject countryrespond to (a) an object country’s specific foreign policy actions and (b) foreignpolicy actions affecting Arab countries other than their own. Unfortunately, existingIR data sets on the dyadic interactions of states are inadequate to test the gener-alizability of this proposition: that is, that majorities of respondents in a givensubject country would alter their view of a given object country were that country toalter its non-military behavior toward another international actor.

Even in early 2002, of course, numerous Arabic and international media sourcescorrectly predicted that the Bush administration would take actions in Iraq thatwould instead exert a downward effect on Arab opinion. In particular, America’searly labeling of Iraq as a member of the ‘‘axis of evil’’ in preparing for a pre-emptive war not contingent on UN support elicited a negative reaction from Arabsfar more severe than that which accompanied the 1991 Gulf War. A not atypicalSaudi commentator, Fawaz Turki, suggested that

Pre-emptive war, such as the one contemplated by the U.S. today, however, is of adifferent order [than the 1991 Gulf War], bringing with it no rhyme and noreason, and fraught as it is with dangerous uncertainties. Moreover, it poses aclear and present danger to all Arabs, in the sense that, collaterally if not directly,an attack on Iraq will be an attack against them all (Turki 2002).

15 The perceived inconsistency of U.S. policies toward Palestine and Iraq probably intensifies animosity toward the

U.S. In 2002, the perception that the U.S. supported Israel at the expense of the rights of the Palestinians wascompounded by a feeling that a double standard existed in its application of international law and justice towardIraq. Arabs pointed out that while Israel occupied the Palestinian West Bank and Gaza for over 30 years, Iraq’sinvasion of Kuwait was quickly condemned and reversed militarily. Similarly, the post-War Iraqi regime was heldunder U.S.-directed UN sanctions, whereas Israel has routinely been sheltered from UNSC actions by U.S. veto.


Predictably then, more recent surveys have found that in the wake of the U.S.invasion of Iraq, negative perceptions of the U.S. have only intensified. Forexample, the Pew Research Center found that in March 2004, only 5% ofJordanians surveyed held a favorable view of the U.S., down from 34% inZogby’s April 2002 survey (Pew Global Attitudes Project, 2004:6; see also Telhami2003b).

If a bit less extensively written about than the salience of Palestine for Arab opin-ion, the salience of Iraq is nonetheless frequently noted by regional experts. Ac-cording to Marc Lynch, for example, ‘‘by virtue of the ongoing debates in theArabist public sphere, Iraq stands as one of the few issues about which it canlegitimately be said that an Arab public opinion exists’’ (Lynch 2003). Table 5certainly suggests that, like the well-being of Palestinians, the well-being of Iraqismay be sufficiently important to Arabs throughout the region to influence theirattitudes toward foreign countries that they see as only secondarily responsible foraffecting that well-being.

It may thus be that Britain’s considerable support for U.S. policies in regard toboth Palestine and Iraq does much to explain Zogby’s respondents ranking of theUnited Kingdom as the third least well-regarded object country in the sample. Britishpolicies in the region are widely discussed in the Arab public sphere, and to an evengreater extent than does Britain’s own press, the Arab press often portrays Britainas a ‘‘lapdog’’ of the United States (see, e.g., Allain 2002). In regard to Iraq inparticular, various 2002 media accounts portrayed Britain as beholden to the U.S.in terms of the maintenance of the Iraq sanctions regime, that is, even beforeBritish support for U.S. policies was reinforced with its role in the 2003 invasion(Hammoud 2002). In regard to Israel, while it is important to remember thatordinary Arabs regard Britain much more favorably than they do Israel itself, itmay also be that Britain’s former imperial role in the region (and in Israel’s cre-ation) continues to resonate for some:

For politicians and intellectuals throughout the Arab world, Israel (or as theyusually called it, ‘‘The Zionist entity’’) was only the reflection of larger and moresinister forces. In their minds, the whole Zionist enterprise was from the begin-ning no more than a facade for British imperialism and then (after Britain’sretreat from the Middle East in the mid-fifties) its American successor (Humph-ries 1999:51).

In short, we suspect that many Arabs perceive Britain as continuing to play a mostlynegative military and political role in the Middle East, albeit a secondary one sinceending its formal colonial role in the region.

Extremely favorable Arab attitudes toward France, in contrast, lead us to rejectany notion that ordinary Arabs are systematically resentful of former imperialpowers. Beyond the fact that, as Table 3 indicates, France is the most favorablyregarded object-country out of all 13, France is nowhere better regarded than in

TABLE 5. Actual and Potentially Favorable Views of the U.S. in Percentages

Subject Country Favorable toward U.S. If U.S. Were to End Sanctions on Iraq

Egypt 15 80Kuwait 41 56Lebanon 26 75Saudi Arabia 12 77UAE 11 62

Table 3; Zogby (2002b).


Lebanon and Morocco (i.e., its two former colonies in the sample).16 This againsuggests to us that Arab attitudes toward particular object countries are stronglyinfluenced by those countries’ relatively recent foreign policy actions in regard toregionally salient issues. In regard to Iraq, for example, French diplomatic pressureto reform the UN sanctions regime had been ongoing for at least 3 years at the timeof the Zogby survey. Similarly, in regard to the Israeli–Palestinian conflict, Francehas, in recent years, taken various diplomatic actions that have led partisans on bothsides of the conflict to view it as the Western country most friendly to Palestine andleast friendly to Israel. Beginning in 2001, for example, France began to (unsuc-cessfully) lobby other EU members to suspend Israel’s EU trade preferences (Black2001).17 In theory, of course, France might have gone beyond merely ‘‘diplomatic’’actions to protest events in the Middle East, but in practice, both Zogby’s findingsand various media accounts suggest to us that many Arabs are appreciative thatFrance has done what it has. French pressure on Israel has, moreover, continuedsince 2002, and positive Arab attitudes toward France continue to be reflected inmore recent surveys (see, e.g., Center for Strategic Studies, University of Jordan2005:19–44).

The second most favorably regarded object country in the Zogby study is Iran,which receives an overall AFFECT rating only one-half a percentage point lowerthan that of France. Viewed in terms of the military, political, economic, and cul-tural characteristics measured in our ‘‘general’’ hypotheses, France and Iran couldhardly be less alike. Yet viewed in terms of their specific foreign policy behavior inthe regionFespecially in regard to Palestine and IraqFboth are, with some jus-tification, perceived as having an active foreign policy in the region distinct from(and sometimes directly in opposition to) the policies of the U.S. and Israel. Wewould thus posit that Iran scored high on AFFECT in large part due to its par-ticularly strident rejection of American and Israeli policies in the region. We stress,however, that we are not arguing that an object country’s foreign policy behavior isthe only contextual factor that influences AFFECT. For example, we suspect thatIran differs from other object countries in that many Arabs may be just as interestedin this neighboring state’s domestic political system as in its foreign policy behav-ior.18 That said, one need hardly posit that object-country foreign policies explainall of the variation in AFFECT in order to improve on existing theory.

With mean AFFECT ratings just above the midpoint at 50, Japan and China seemcomparable as two major powers that are not deeply involved in the geopolitics ofthe Middle East. Although both are consumers of Middle Eastern oil, we found noArab media depictions of either country as projecting military power or interferingpolitically in the region. Although in the Arab public sphere there is some recog-nition of the potential hegemony of Japan and the rising military power of China,neither country’s military power is often addressed. Rather, most media coverage ofthe two countries centers on issues of economics and trade, with the successfuleconomic development of the two states sometimes cited as a model for Arab de-velopment (Eisa 2002; Elewa 2002; Nafie 2002). To many of the Arabs surveyed,moreover, we suspect that the two countries are simply unfamiliar, a possibility that

16 Surveys were not conducted in other former French colonies like Algeria, Syria, and Tunisia, which may ormay not share Lebanon and Morocco’s enthusiasm for France.

17 This suggests to us that an object country’s diplomatic behavior may be a generalizable, if not easily oper-ationalizable, predictor of attitudes toward that country abroad.

18 Although Iran is the only Muslim object country that Arab publics view favorably, we suspect that some Arabsview it as a successful experiment in the blending of Islamic values and popular participation in government. The

Zogby data give no indication of how the predominantly Sunni Arabs polled feel about Persian Shi’ism, but they dosuggest that the vast majority of respondents endorse political participation (Zogby, 2002a:34). Iran’s fairly recentconflicts with Arabs during the Iran–Iraq War are also potentially relevant (and may help to explain why Iran’s morestrident opposition to U.S. and Israeli policies has not rendered Iran even more popular than France). Again,however, we do not claim that an object country’s foreign policy behavior is the only influence on AFFECT.


seems consistent with the high degree of variation in attitudes toward both (SD forJapan ¼ 14.4; SD for China ¼ 16.7).

Moving now to countries with mean AFFECT ratings just below the midpoint at50, it may be that Germany and Canada are seen as U.S. allies that nonethelesspromote development and human rights in the region and limit their militarypresence to peacekeeping operations (Canada) and training exercises (Germany).Outside of the UAE, the majority of respondents to the Zogby survey view Germanyand Canada favorably. As with China and Japan, however, the high degree ofvariability in AFFECT ratings for these two countries (SD for Germany ¼ 16.2; SDfor Canada ¼ 16.4), suggests that they are influenced not only by considered re-spondent neutrality but also by low levels of respondent information.

That Arab attitudes toward Pakistan (AFFECT ¼ 46) and India (AFFECT ¼ 41)are roughly comparable is consistent with our failure to find support for Hunt-ington’s hypotheses about the civilizational determinants of amity and enmity. Canthe similarly ‘‘cool’’ ratings that these two South Asian rivals receive thus be partlyexplained in terms of their foreign policy behaviors in the region? Although neitherPakistan nor India is deeply involved in the Middle East, the foreign policies of bothhave been criticized in the Arab press. In particular, Pakistan’s frequent militarycooperation with the U.S. has not gone unnoticed (Alam 2002; Haqqani 2002).India, for its part, is perceived by some commentators as increasing its military tiesto Israel, though others also note its uneven treatment of Indian Muslims (Tash2001; Masrawah 2002).19

The object country immediately below India in terms of favorable attitudes isRussia (AFFECT ¼ 38). Russia’s role in the Middle East is written about fairly ex-tensively, with various commentators discussing it as a potential international rivalof Arab states (Nafaa 2001), as a victim of Russian pro-Israeli lobbies (Naguib 2002),and as a co-conspirator with American foreign policy (Seale 2001). Although not allof these arguments seem plausible, the merely cool feelings toward Russia foundamong Arab respondents do seem consistent with its mixed foreign policy record inthe region. Russia is likely resented for its membership in the unpopular ‘‘Quartet’’of powersFseen by many Arabs as dictating a pro-Israel resolution of the Israeli–Palestinian conflictFbut we also suspect that many Arabs are aware that Russiancooperation with U.S. policies in regard to Iraq has been less enthusiastic than thatof countries such as the United Kingdom.

Finally, respondent attitudes toward Turkey (AFFECT ¼ 34) are almost as cool asthose toward the U.K. The fact that Turkey has significant military ties to Israel iswidely criticized in the Arab media, and in early 2002, some commentators linkedTurkey’s cooperation with Israel to its initial cooperation with the U.S. in prep-aration for a war against Iraq (Nassar 2002). Press coverage of the Turkish par-liament’s later rejection of U.S. proposals to base troops for the Iraq War in Turkeywas more positive, but we lack more recent survey data with which to explore thepossibility that this rejection has led to a softening of Arab attitudes about Turkey.

A Middle-Range Theory of Arab Opinion

The idea that publics evaluate foreign countries largely on the basis of relativelyrecent and specific foreign policy behaviors is not entirely unanticipated by pre-vious IR theorists. Hans Morgenthau, for example, was perhaps the first to arguethat foreign policy opinion is ultimately driven by ‘‘concrete’’ concerns: Like Waltz,Morgenthau assumed that the ‘‘abstract’’ foreign policy attitudes of publics werepacific. Unlike Waltz, however (who concluded that public pacifism was little af-fected by international events and, in turn, analytically unimportant), Morgenthau

19 To our knowledge, neither Zogby nor other survey researchers have asked Arabs about their sympathies withforeign Muslims other than the Palestinians (e.g., those in Kashmir and Chechnya).


suggested that this pacifism quickly turns to popular nationalism and militarismwhen an enemy threatens a country’s interests.

When actual war threatens in our time. . . humanity remains united in its horrorof war as such, and in its opposition to it. But men are incapable of translating thisabstract opposition to war as such into concrete opposition to this particular war.While most members of the human race, qua members of the human race, con-sider war under the conditions of the mid-twentieth century an evil. . . mostmembers of the human race, qua Americans, Chinese, Englishmen and Russians,look at war as they have always done, from the point of view of their particularnations (Morgenthau 1954:242).

For Morgenthau, in other words, public attitudes toward particular foreign coun-tries are mostly determined by the question, ‘‘what has this country done for (or to)my own country lately?’’ If a bit difficult to operationalize, moreover, it is certainlypossible to state another ‘‘general’’ hypothesis broadly consistent with Mor-genthau’s arguments: that is, ‘‘citizens of Country A will evaluate Country B on thebasis of Country B’s specific foreign policy actions (and especially their militaryactions) in regard to Country A.’’

Whatever the promise of such a hypothesis in other contexts, however, we thinkit requires considerable modification if it is to be predictive of Arab opinion onforeign relations. Specifically, the hypothesis that we now propose suggests thatArabs in a given subject country will be concerned not only with object-countrymilitary behavior toward their own country, but also with a wider range of objectcountry behavior affecting Arabs throughout the region.

Hypothesis 7: Arab publics evaluate object countries based on the military and non-militarypolicy behaviors of those countries in regard to regionally salient issues.

A systematic evaluation of this hypothesis will, among other things, require thecollection of new data other than that upon which it has been formulated. Thehypothesis will certainly appear more promising, however, insofar as ordinary Ar-abs’ identifications and allegiances can be shown to extend beyond the individualcountries in which they reside. Although there is no single Arab public opinionabout foreign relations, we have seen that all seven Arab publics surveyed expressstrong concerns about certain issuesFfor example, Palestine and IraqFthat do notaffect their own country’s ‘‘national’’ interests narrowly defined.20 In our view, thishigh level of concern with ‘‘regional’’ matters is best understood within the par-ticular historical context of the twentieth and early twenty-first century Middle East.

Arab Opinion in the Context of Competing Identities

Since the decline and fall of the Ottoman Empire, the Arab public sphere hasfamously been the site of multiple attempts to define the scope of the nation and itsrelationship to state boundaries. Widely shared historical legacies (of colonialism,independence, the Cold War, and oil economics) have left public opinion in manycountries in the region in a domestic context characterized by a competition be-tween multiple frames of identity. Today, three major identity frames compete todefine the proper scope of the ‘‘nation’’ in the Arab world: Arabism (‘‘qawmiyya’’),Islamic (or more properly Islamist) identity, and ‘‘country’’ specific nationalism

20 As we discuss below, however, there is also considerable subject-country variation in evaluations of particular

countries. In any case, the argument that Arab publics tend to share perceptions about key foreign policy issues isentirely distinct from any vulgar ‘‘Arab Street’’ thesis, suggesting that Arab public opinion is monolithically anti-Western or hostile to outsiders (Lewis 2002; Pipes and Schanzer 2002). Rather, in rejecting Hypothesis 6, we havealready rejected the claim that Arab opinion is monolithically anti-Western. (A quick glance at Table 3 suffices todemonstrate that Arabs are not systematically hostile to foreign countries in general.)


(‘‘wattaniyya’’). Each of the three seeks to reconcile the effects of the colonial divisionof the Arab world, the lingering divisions in the Muslim community, and the con-temporary norms of Westphalian sovereignty. Arabism and Islamism seek to sub-sume the colonially drawn boundaries in a larger community, while country-specificnationalism takes the current map as a given and seeks to build a community withinexisting states.21

Competition between these three logics of organizing the ‘‘national’’ communityin Arab societies often seems to lie just below the surface of public debate on foreignrelations. In turn, how individual respondents surveyed by Zogby define theiridentity may be important in explaining their evaluations of particular objectcountries. Given space constraints and the present unavailability of Zogby’s indi-vidual-level data, we cannot yet test this inference in a rigorous way.22 What we cando is note that Zogby’s societal-level data on identity is broadly consistent with thefact that respondents seem to care a great deal about foreign policy behaviorsaffecting Arabs outside their own countries. Specifically, Zogby finds that respond-ents generally ranked ‘‘Arab’’ as their most important category of personal iden-tification (mean: 41%) over ‘‘religion’’ (mean: 24%), citizen of a particular country(mean: 18%), and various sub-national forms of identification (combined mean17%) (Zogby 2002:49–52). The weakness of ‘‘nation-state’’ identity relative to ‘‘Ar-abist’’ identity strikes us as particularly remarkable.23

Of course, to argue that allegiance to the nation state is not the primary mode ofallegiance among Zogby’s respondents is not to suggest that the publics of differentArab countries fail to vary in their attitudes toward particular international actors.On the contrary, we now conclude our empirical analysis of object-country foreignpolicy behavior as a determinant of Arab opinion by examining opinion in threeindividual subject countries (Morocco, Lebanon, and Kuwait) in which relativelyunique historical legacies and international experiences seem particularly impor-tant determinants of amity and enmity.

Country-Specific Variation in Arab Foreign Policy Opinion

Morocco’s geographic and, to some extent, historical isolation from the other Arabcountries surveyed deserves attention. It is important to note that Morocco was theonly country surveyed that was not historically part of the Ottoman Empire. Thus,while Moroccans still see Palestine as an important issue (92%), this perception canstand alongside relatively favorable views toward Israel, the U.S. and Turkey; Mo-roccans hold the most favorable views toward France, Israel, Turkey, and Germany,with greater than a standard deviation above the mean of all subject country eval-uations. Evaluations of the U.S. were also nearly a standard deviation above themean and second only to favorable Kuwaiti evaluations. We presume that manyhistorical conflicts in the central Middle East are not as salient for the Moroccanpublic as for those publics that were more involved in them. In short, the substan-tial, though not overwhelming, impact of varied historical experiences of differentArab states is well exemplified in Moroccan attitudes toward numerous non-Arabcountries.

The legacies of country-specific histories are likewise apparent in attitudes onforeign relations among the Lebanese publicFattitudes which appear affected by

21 We discuss the salience of various sub-state identities with family, tribe, and region of country below, but thesemodes of identification are less pertinent to our present examination of attitudes on foreign relations.

22 We intend to analyze Zogby’s still-embargoed individual-level data in the future.23 While the strength of Arabist identity in the general population may be somewhat exaggerated by Zogby’s

oversampling of ‘‘attentive’’ publics, evidence from the World Values Surveys (Inglehart 2001) suggests that suchoversampling would also exaggerate ‘‘nation-state’’ identification (i.e., due to the fact that the Third World ruralpoor are systematically likely to express ‘‘sub-national’’ allegiances). In turn, vagaries of the Zogby sample wouldappear to have little influence on the relative strength of ‘‘Arabism’’ and ‘‘nation-state’’ identification.


the bitter and quite recent experience of that country’s civil war. Table 4 indicatesthat Lebanon is the only country in which fewer than 80% of respondents to theZogby survey rate the issues of ‘‘Palestine’’ (69%) and the ‘‘rights of Palestinians’’(67%) as ‘‘important’’ or ‘‘very important.’’ This lesser concern with Palestine isalmost surely related to the involvement of Palestinians in the 15-year civil war and,more specifically, to negative perceptions of Palestinian refugees in Lebanon (Fisk1992; Collings 1994). Logically enough, however, diminished sympathy for Pales-tinians by no means translates into favorable Lebanese evaluations of Israel(AFFECT ¼ 5%), which occupied parts of Lebanon from 1978 until 2000.

A third notable case of country-specific variation concerns the considerably morefavorable views of the U.S. and U.K. held by respondents in Kuwait. Favorable andvery favorable perceptions of the U.S. and Britain among Kuwaitis are roughly 16%higher and more than one standard deviation above the overall means. Here again,it seems uncontroversial to assert that American and British leadership in the 1991Gulf War, the subsequent enforcement of sanctions against Iraq, and looming plansfor a second invasion had a different meaning in Kuwait than in much of the rest ofthe Arab world. Kuwaiti exceptionalism should not, however, be overstated. AsLynch points out, ‘‘while the memory of 1990–1991 was doubtless sufficient toinsure Kuwaiti hostility toward Iraq, Kuwaitis nonetheless recognized the shiftingterrain of the Arab consensus [on Iraq]’’ (Lynch 2003:80). Indeed, even 15% ofKuwaiti respondents saw the sanctions on Iraq as a decisive factor in their negativeevaluation of the U.S. (see Table 5).

In sum, particular national contexts and histories of dyadic relationships, do, asmany traditional IR realists might predict, impact patterns of amity and enmity inparticular dyads.24 What no IR theoristFand certainly no IR realistFseems yet tohave anticipated, however, is that survey respondents in multiple Arab countrieswould express such deep concern with object-country behaviors affecting persons whoare not their compatriots. In the conclusion of our paper, we explore the implicationsof such concern for theory development on the one hand and for policy on the other.

Conclusion

General theories of international relations would seemingly offer great help inplacing rhetorically charged debates about Arab public opinion in a rigorous social-scientific context. As it turns out, however, this help is entirely negative: empiricaltests of existing Realist, Liberal, Marxist, and Cultural hypotheses fail to signifi-cantly explain the roots of Arab public opinion toward the 13 object countries in thestudy. On the contrary, we have presented inductive evidence suggesting that Arabpublics evaluate non-Arab object countries on the basis of those countries’ specificforeign policy behaviors throughout the wider Middle East. Otherwise dissimilarcountries seen as playing a positive role in the regionFfor example, France andIranFare regarded favorably, whereas dissimilar countries seen as playing a neg-ative roleFfor example, Israel, the U.S., and TurkeyFare not. At present, anobject country’s policy actions in regard to Palestine and Iraq seem to have par-ticular importance in determining how that country is regarded. Treatment ofPalestinians and Iraqis by various members of the international community iswidely covered in the Arab media, and we have presented evidence that Arabs treatthese two issues as ‘‘litmus tests’’ in forming either favorable or unfavorable opin-ions of foreign countries.

24 Even many of these strictly ‘‘dyadic’’ determinants of AFFECTwould require considerable development of IRdata sets before they could be tested quantitatively. For example, it may well be thatFas appears to be the case withthe Kuwait-U.S. dyadFa subject country’s ‘‘gratitude’’ toward an object country for ‘‘aiding in restoration of itsterritorial integrity’’ declines predictably over time. Before testing the generalizability of this hypothesis, however,one would need to quantify ‘‘third-party aid in restoration of subject-country territorial integrity.’’


We have further argued that the widespread concern of Arabs outside Iraq andPalestine with the well-being of Iraqis and Palestinians is consistent with the pri-macy of ‘‘Arabist’’ frames of identity among respondents to the Zogby survey.Although our own analysis has been provisional in this regard, we hope that futureanalysts of Arab attitudes on foreign relations will pay careful attention to thecompeting frames of identity and, in particular, the various ‘‘supra-state’’ frames ofidentity through which such attitudes may be filtered. (The extent to which analystsof amity and enmity in other regions should be similarly attentive to supra-stateidentifications would logically depend on their empirical prevalence in those re-gions.)25 A second and perhaps more generalizable implication of our research isthat all scholars of amity and enmity should pay close attention to a variety of ‘‘non-military’’ state behaviors that have yet to be operationalized quantitatively.Although analyzing such behaviors as predictors of subject country attitudes to-ward particular object countries is further limited by the paucity of cross-nationalsurvey data on such attitudes, trans-regional analyses of these predictors seem aslikely as any to uncover robust ‘‘general’’ determinants of amity and enmity.26

In any case, the fact that we have here found existing theories of amity andenmity unhelpful in predicting Arab attitudes on foreign relations should actuallybe cause for optimism on the part of policy makers. First, and most obviously, ourfindings suggest that countries wishing to improve their image among ordinaryArabs can do so. Expressed in the language of the pundits, we find no evidence thatordinary Arabs resent countries for what they ‘‘are,’’ and considerable evidencethat they resent them for what they ‘‘do.’’ Second, and perhaps more subtly, we findthat countries wishing to improve their image can do so with great efficiency. Spe-cifically, it appears that a state wishing to ‘‘win the hearts and minds’’ of the publicsof, for example, Egypt, Jordan, Lebanon, Kuwait, Morocco, Saudi Arabia, and theUnited Arab Emirates can vastly improve its standing in all of these societies simplyby altering its policies in regard to Palestine and Iraq.

Needless to say, the most salient issues of the present may not be salient in thefuture. Although the issues of Palestine and Iraq have been salient for some time,new historical contingencies may alter the attitudes of ordinary Arabs about theappropriate role in the region of particular countries. Whatever political changesmay occur in the Middle East, however, it seems highly likely that policy makers willcontinue to be able to consult reasonably reliable surveys on how ordinary Arabsperceive them. At present, theory and evidence alike indicate that the way to im-prove one’s country’s image in Arab opinion is not to run roughshod over thatopinion but to heed it.

References

ALAM, ABSAR. (2002) Friends in Deed, Al-Ahram Weekly, April 4. http://weekly.ahram.org.eg/2002/580/in13.htm.

ALLAIN, JEAN. (2002) Criminal Enforcers, Al-Ahram Weekly, May 9. http://weekly.ahram.org.eg/2002/585/op10.htm.

BARNETT, MICHAEL. (1998) Dialogues in Arab Politics: Negotiations in Regional Order. New York: ColumbiaUniversity Press.

BENNETT, D. SCOTT, AND ALLAN STAM. (2000) EUGene: A Conceptual Manual. International Interactions26:179–204.

25 Global data on the salience of ‘‘regional’’ identities suggest that the high levels of ‘‘pan-Arabism’’ observed byZogby fail to be paralleled by high levels of either ‘‘pan-Americanism,’’ ‘‘pan-Asianism,’’ or ‘‘pan-Europeanism.’’ See

among others, Inglehart (2001).26 Although policy-oriented work on the determinants of ‘‘soft power’’ (primarily the soft power of the United

States) has not yet led to global quantitative studies that attempt to operationalize soft power in a comparativecontext, we infer from Joseph Nye’s work that such an operationalization might well resemble our own opera-tionalization of each object country’s ‘‘mean score’’ for AFFECT (see Nye 2004:35–44).


BENNETT JR., D. SCOTT, AND ALLAN C. STAM III. (2003) The Behavioral Origins of War. Ann Arbor:University of Michigan Press.

BLACK, IAN. (2001) Europe Turns Heat on Israel; France Leads Call for Action, The Guardian, March14.

BLAINEY, GEOFFREY. (1973) The Causes of War. New York: Free Press.CENTER FOR STRATEGIC STUDIES, UNIVERSITY OF JORDAN. (2005) Revisiting the Arab Street: Research From

Within. Amman: Center for Strategic Studies.CHICAGO COUNCIL ON FOREIGN RELATIONS. (1995, 1999, 2003) American Public Opinion and US Foreign

Policy 1994, 1998, and 2002 Reports. Chicago: Chicago Council on Foreign Relations.COLLINGS, DEIRDRE, ED. (1994) Peace for Lebanon?: From War to Reconstruction. Boulder: Lynne Rienner

Publishers.DOYLE, MICHAEL. (1983) Kant, Liberal Legacies and Foreign Affairs, Parts 1 and 2. Philosophy and

Public Affairs 12:205–254, 323–353.EISA, MOHAMED ABDEL-SHAFEI’I. (2002) Asian Technology: Does Egypt Need It?, Al Ahram Weekly

Online, May 30. http://weekly.ahram.org.eg/2002/588/ec4.htm.ELEWA, AL-SAYED. (2002) Eastern Experience, Al-Ahram Weekly Online, January 31. http://weekly.ah-

ram.org.eg/2002/571/op7.htm.FISK, ROBERT. (1992) Pity the Nation: Lebanon at War. 2nd edition. Oxford: Oxford University Press.FRIEDMAN, THOMAS. (2002a) Wall of Ideas, The New York Times, March 3.FRIEDMAN, THOMAS. (2002b) The Hidden Victims, The New York Times, May 2.FRIEDMAN, THOMAS. (2003) A Theory of Everything, The New York Times, June 1.FURIA, PETER. (1999) Public Opinion and the Democratic Peace. A Paper presented at the annual

meeting of the American Political Science Association, Atlanta, GA.GILPIN, ROBERT. (1981) War and Change in World Politics. Cambridge: Cambridge University Press.HAMMOUD, HAMID. (2002) The US and the Iraqi Regime. Al-Hayat (London), February 22. http://

english.daralhayat.com/opinion/02-2003/Article-20030222-85e4ffc7-c0a8-01fc-0015-2643ac1884b7/story.html.

HAQQANI, HUSAIN. (2002) Rules Made Arbitrarily Do Not Bring Stability. Gulf News, June 30. http://www.gulf-news.com/Articles/print2.asp?ArticleID=56022.

HOLSTI, OLE. (1996) Public Opinion and American Foreign Policy. Ann Arbor: Michigan University Press.HUMPHRIES, R. STEPHEN. (1999) Between Memory and Desire: The Middle East in a Troubled Age. Berkeley:

University of California Press.HUNTINGTON, SAMUEL. (1993) The Clash of Civilizations? Foreign Affairs 72:22–50.HUNTINGTON, SAMUEL. (1996) The Clash of Civilizations and the Remaking of World Order. New York:

Simon & Schuster.INGLEHART, RONALD. (2001) World Values Surveys and European Values Surveys, 1981–1984, 1990–1993

and 1995–1997. Ann Arbor: Inter-University Consortium for Political and Social Research. StudyNo. 2790.

INTERNATIONAL MONETARY FUND. (2002) Direction of Trade Statistics Yearbook. Washington: InternationalMonetary Fund.

JONES, D. M., S. A. BREMER, AND J. D. SINGER. (1996) Militarized Interstate Disputes, 1816–1992. AnnArbor: Inter-University Consortium for Political and Social Research.

JOFFE, JOSEF. (2003) Anti-Americanism, Anti-Israelism, Anti-Semitism: Do the Three Travel Together,and Why, Public Lecture at Hebrew University of Jerusalem, March 6.

KAGAN, ROBERT. (2003) Of Paradise and Power: America and Europe in the New World Order. New York:Alfred A. Knopf.

KERR, MALCOM H. (1971) The Arab Cold War: Gamal Abd al-Nasir and His Rivals, 1958–1970. London:Oxford University Press.

KIFNER, JOHN. (2001) The New Power of Arab Public Opinion, The New York Times, November 11.KISSINGER, HENRY. (1994) Diplomacy. New York: Simon & Schuster.KHALIDI, RASHID. (2002) Powell Must Seek Real Compromise, The Washington Post, April 11.KHASHOGGI, JAMAL. (2002) It’s the Occupation, Stupid, Arab View, April 3. http://www.arabview.com/

articles.asp?article=106.KORANY, BAHGAT, PAUL NOBLE, AND REX BRYNEN EDS. (1993) The Many Faces of National Security in the

Arab World. New York: St. Martin’s Press.KRAUTHAMMER, CHARLES. (2002) Where Power Talks, The Washington Post, January 4.KRISTOF, NICHOLAS. (2002) Why Do They Hate Us?, The New York Times, January 15.LEWIS, BERNARD. (2002) What Went Wrong? The Atlantic Monthly 289:43–46.LYNCH, MARC. (2003) Beyond the Arab Street: Iraq and the Arab Public Sphere. Politics and Society

31:55–91.


MASRAWAH, TARIQ. (2002) The Indian-Israeli Threat, Al-Quds Al-Arabi (London), January 21, reprintedin Mideast Mirror, January 22, 2002.

MEARSHEIMER, JOHN. (1990) Back to the Future. International Security 15:5–56.MORGENTHAU, HANS J. (1954) Politics among Nations: The Struggle for Power and Peace. 2nd edition.

Revised and enlarged. New York: Alfred A Knopf.NAFAA, HASSAN. (2001) Trying to Spot the Difference, Al Ahram Weekly Online, December 27. http://

weekly.ahram.org.eg/2001/566/op12.htm.NAFIE, IBRAHIM. (2002) Perils and Opportunities, Al-Ahram Weekly Online, January 17. http://weekly.

ahram.org.eg/2002/569/op1.htm.NAGUIB, SHOHDY. (2002) Deflecting the Truth, Al-Ahram Weekly Online, May 2. http://weekly.ahram.

org.eg/2002/584/11inv2.htm.NASSAR, GALAL. (2002) The Axis of Evil–from Another Angle, Al-Ahram Weekly Online, March 7. http://

weekly.ahram.org.eg/2002/576/focus.htm.NYE, JOSEPH. (2004) Soft Power: The Means to Success in World Politics. Cambridge: Perseus.ORGANSKI, A. F. K., AND J. KUGLER. (1980) The War Ledger. Chicago: University of Chicago Press.PEW GLOBAL ATTITUDES PROJECT. (2004) A Year After the Iraq War: Mistrust of America in Europe Even

Higher, Muslim Anger Persists. Washington, DC: Pew Research Center for the People and the Press.PIPES, DANIEL, AND JONATHAN SCHANZER. (2002) Muslim World Denial is Cause for Disaster, Chicago

Sun-Times, January 20: 33.ROGER, PHILIPPE. (2002) L’ennemi Americain. Paris: Seuill.RUSSETT, BRUCE. (1993) Grasping the Democratic Peace. Princeton, NJ: Princeton University Press.RUSSETT, BRUCE, AND JOHN ONEAL. (2001) Triangulating Peace. New York: Norton.SARDAR, ZIAUDDIN, AND MERRYL WYN DAVIES. (2002) Why Do People Hate America? New York: Disin-

formation.SEALE, PATRICK. (2001) Will the Arabs Pay the Price of the Bush–Putin Honeymoon? Al-Hayat (London),

November 27, reprinted in Mideast Mirror, November 28, 2001.SINGER, J. DAVID, STUART BREMER, AND JOHN STUCKEY. (1972) Capability Distribution, Uncertainty, and

Major Power War, 1820–1965. In Peace, War and Numbers, edited by Bruce Russett. Beverly Hills:Sage.

SULFARO, VALERIE, AND MARK CRISLIP. (1997) How Americans Perceive Foreign Policy Threats: AMagnitude Scaling Analysis. Political Psychology 18:103–126.

TASH, ABDUL QADER. (2001) Lies About India’s Stand on Palestine, Arab View, August 23. http://www.arabview.com/articles.asp?article=92.

TELHAMI, SHIBLEY. (2001) Sympathy for the Palestinians, The Washington Post, July 25.TELHAMI, SHIBLEY. (2002) The Stakes: America and the Middle East. Boulder: Westview Press.TELHAMI, SHIBLEY. (2003a) Is Iraq More Crucial to Regional Peace Than Palestine? An Arab-Israeli

Pact Must Come First, The New York Times, April 12.TELHAMI, SHIBLEY. (2003b) Arab Public Opinion on the United States and Iraq: Postwar Prospects for

Changing Prewar Views. Brookings Review 21:24–27.TESSLER, MARK, AND JODI NACHTWEY. (1998) Islam and Attitudes Toward International Conflict. Jour-

nal of Conflict Resolution 42:619–636.TURKI, FAWAZ. (2002) Washington’s Obsession with Iraq, Arab News, January 8. http://www.arabview.

com/articles.asp?article=142.WALLERSTEIN, IMMANUEL. (2002) The Eagle Has Crash-Landed. Foreign Policy 131:60–68.WALTZ, KENNETH. (1959) Man, the State and War. New York: Columbia University Press.WATERBURY, JOHN. (2003) Hate Your Policies, Love Your Institutions. Foreign Affairs 82:58.WRIGHT, ROBERT. (2002) A Real War Against Terrorism, Slate, September 11.ZOGBY, JAMES. (2002) What Arabs Think: Values Beliefs and Concerns. Utica, NY: Zogby International.ZOGBY INTERNATIONAL. (2002) The Ten Nation Impressions of America Poll. Utica, NY: Zogby Interna-

tional.


606

Determinants of Arab Public Opinion on Foreign Relations

Documents